Six months later, no one uses it
A team in a mid-sized company buys a robotic process automation (RPA) platform. They have a clear target: a set of legacy desktop applications that still rely on manual data entry. Six months later the bots are running, but the expected time savings have not materialized. The invoices still get entered twice. The error rate barely budged. The automation team is blamed for picking the wrong tool. But the real mistake was picking the wrong category.
I have watched this pattern more times than I can count. McKinsey tracked it and found that 73% of failed automation projects happen because teams automate a broken process instead of fixing it first. That figure comes from Stepper.io citing McKinsey. The source sells BPA software, so I treat the number as a strong directional signal—but it aligns with what I see every time: the problem is not vendor choice, it is category choice.
The market offers five distinct categories: RPA, BPA, iPaaS, workflow automation, and AI-agent platforms. Each solves a different kind of problem. The process automation tool you end up choosing should be the last thing you decide, not the first. Here is a quick litmus test for each. Read the question. If it matches the process you are trying to automate, that category is your starting point.
| Category | What it automates | Litmus-test question | Best for |
|---|---|---|---|
| RPA | Repetitive desktop tasks in legacy applications | Are you clicking the same buttons in a Windows app every day? | Finance, HR, admin – any team stuck with green‑screen or thick‑client apps |
| BPA | Cross-department business processes with human handoffs | Does the process involve approvals, forms, and multiple people doing steps in sequence? | Operations, compliance, loan processing – processes with human decision points |
| iPaaS | Real-time data sync and complex integrations between cloud apps | Do you need to move data between Salesforce, ERP, and a data warehouse without writing glue code? | Data teams, IT – any scenario requiring many API calls and transformations |
| Workflow Automation | Connecting SaaS apps with linear, no‑code steps | Are your tools web‑based and do you just need to pass data from one app to the next? | Solo operators, small teams, marketing, sales – common SaaS stack |
| AI Agents | Autonomous tasks using LLMs, logic, and external tools | Do you need the system to decide what to do next based on context and previous outputs? | Innovation teams exploring generative AI; advanced orchestration beyond fixed rules |
Gartner predicts more than 40% of agentic AI projects will be canceled by end of 2027 – cited on the Blue Prism blog. The stated reasons are escalating costs, unclear business value, or inadequate risk controls. That is not a reason to avoid AI agents – it is a reason to be sure you need one before you start. If a plain workflow automation tool can do the job, use it.
Which category fits your situation?
The litmus tests get you most of the way. Your own context – technical skill, budget, data-residency needs, scale – can push you toward one category over another even when the process seems to fit two. Here is a decision matrix that matches reader profiles to best-fit categories.
For example, if you are a non‑technical solo operator with a web‑based tool stack, workflow automation is almost always the right first category. If you are in IT Ops and need to orchestrate on‑prem systems alongside cloud apps, you may need a BPA or iPaaS tool – 88% of enterprises run hybrid IT, per a Stonebranch report. That number comes from a company that sells IT automation, so I treat it as a strong signal rather than a precise measure. Still, it tells me most readers are not at pure-cloud scale. Options like n8n (self‑hosted) or ProcessMaker cover that hybrid need.
If you are still unsure, the Workflow Orchestration vs Automation: The 4-Question Litmus Test goes deeper into the difference between iPaaS and simple workflow automation. And for non‑technical teams, Process Automation Tools for Non-Technical Teams covers exactly which platforms actually work without a developer.
Comparing tools within your chosen category
Once you know the category, the real tool comparison begins. I will walk through the three most common categories people land on – workflow automation, AI agents, and iPaaS. For RPA and BPA, the tool landscape is narrower, so I will keep those brief.
Workflow automation: Zapier, Make, n8n
| Tool | Integrations | Starting price (billed annually) | Scale ceiling | Best fit |
|---|---|---|---|---|
| Zapier | 6,000+ | $19.99/mo (750 tasks) | Gets expensive at volume; at 100K ops/mo >$300 | Solo ops, simple linear automations, huge app selection |
| Make | 1,500+ | $9/mo (10K operations) | Good value at scale: at 100K ops/mo <$100 | Data‑heavy workflows, complex logic, small teams |
| n8n | 1,000+ | Free (self‑hosted); cloud $22/mo (2.5K executions) | Self‑hosted: zero per‑execution cost; cloud scales well | Compliance‑minded teams, developers, hybrid IT |
For a deeper dive into these three specifically, read Zapier vs Make vs n8n: Why Choosing the Wrong Process Automation Tool Costs You Time and Money.
AI agents: n8n, Zapier Agents, Make Maia
| Tool | AI capability | Integrations | Pricing model | Best fit |
|---|---|---|---|---|
| n8n | LangChain integration (70+ AI nodes) | 1,000+ | Free self‑hosted / cloud $22/mo | Teams that need control over LLM usage and data privacy |
| Zapier Agents | Autonomous task execution across 8,000+ apps | 8,000+ | Per‑task pricing (same as core plan) | Quick AI experiments, non‑technical users |
| Make Maia | AI assistant that builds scenarios from natural language | 1,500+ | Per‑operation pricing (starts $9/mo) | Users who want to prototype AI workflows without coding |
For a more technical comparison of the AI‑agent architectures behind these tools, see Process Automation Tools Compared: RPA vs. No-Code Workflow vs. AI Agents — Which Architecture Wins in 2026?. That piece targets technical leads; the framing here – taxonomy‑first – is the one that matters for non‑technical decision‑makers.
iPaaS: Workato, Boomi (and why most readers do not need them)
iPaaS tools like Workato (2,000+ integrations, supports batch actions) and Boomi are built for enterprise‑scale integration – real‑time sync between ERP, CRM, and data warehouses. They come with enterprise pricing and corresponding complexity. If you are a solo operator or a small team, you probably do not need an iPaaS. The Stonebranch report found that only 21% of organizations have reached enterprise‑scale AI workflow deployment; the gap is orchestration, not capability. For the other 79%, a workflow automation tool or a self‑hosted n8n is usually sufficient.
RPA: UiPath, Automation Anywhere (and a caveat)
RPA is a $28.31B market (Precedence Research 2025) and it is mature. It is the right choice if you are automating legacy desktop applications that have no API. But many teams buy RPA when what they actually need is workflow automation for web apps. Use the litmus test: if you are clicking inside a Windows application, go RPA. If you are connecting web apps, go workflow automation.
BPA: Kissflow, ProcessMaker, and the on‑prem need
BPA tools are for processes with human approvals, forms, and cross‑department handoffs. The global BPA market is projected to grow from $22.45B in 2026 to $54.34B by 2032 (Fortune Business Insights, cited by Stepper.io). That number comes from a source that sells BPA software, so I treat it as a directional signal – but the rapid growth story aligns with what I see in practice. If your process involves a manager reviewing a purchase order, a finance person checking a budget, and an admin finalizing a report, BPA is your category. Tools like Kissflow and ProcessMaker cover this space well. For on‑prem needs (88% of enterprises run hybrid IT), ProcessMaker offers a self‑hosted option.
- Best for solo ops: Zapier — huge app library, lowest learning curve.
- Best for data‑heavy workflows: Make — per‑operation pricing makes it far cheaper than Zapier at volume.
- Best for compliance‑minded teams: n8n self‑hosted — zero per‑execution cost, full data control.
- Best for Microsoft shops: Power Automate — deep Office 365 integration, already in the ecosystem.
- Best for enterprise orchestration: Workato — batch actions reduce costs, strong API management.
- Best for desktop legacy apps: UiPath — the RPA leader.
FAQs: Can I migrate between categories later?
The Stonebranch report tells us that 89% of enterprises now manage multiple automation platforms. That is not a sign of failure – it is the normal state of an organization that automates different kinds of processes. You will probably end up with more than one tool over time.
If you are trying to decide whether you have outgrown simple workflow automation and need orchestration or iPaaS, the Workflow Orchestration vs Automation: The 4-Question Litmus Test will give you a clear answer.
Start with the category. The tool will follow.